search

MCZhi / DIPP

[TNNLS] Differentiable Integrated Prediction and Planning Framework for Urban Autonomous Driving
https://mczhi.github.io/DIPP/
197 stars 40 forks source link

Training error: torch._C._LinAlgError: torch.linalg_cholesky #6

Closed ydcttt closed 1 year ago

ydcttt commented 1 year ago

Hi, this is excellent work. I try to reproduce it, but when training at Epoch 6/20, I got this error torch._C._LinAlgError: torch.linalg_cholesky: (Batch element 11): The factorization could not be completed because the input is not positive-definite (the leading minor of order 48 is not positive-definite). I want to ask how to solve it. image

MCZhi commented 1 year ago

Please see #1. This issue happens sometimes if using a dense solver, so I suggest using a sparse solver instead.